Building a product in 48 hours and hitting $30k MRR as a non-technical founder

Hasaam Bhatti is a non-technical founder who built a product in 48 hours and hit $10k MRR in 30 days. A few months later, Launch Fast is at $30k MRR

Here's Hasaam on how he did it. 👇

Building a SaaS in 48 hours

I worked a corporate job and ran two Amazon brands on the side when I found Cursor on Twitter. No CS degree. Never wrote a line of code,

I'd tried building products before, AI video tools, job automation apps. Ten or twelve of them. None made it to production because I was building for audiences I didn't belong to.

Amazon was different. I spent 20 to 30 hours researching every product idea, copy-pasting data into Google Sheets. I was frustrated, and I knew exactly what was broken. So I built the tool I needed.

Distribution was the harder problem. Zero audience means zero customers. But I'd bought a coaching program called Legacy X two years earlier, and it had thousands of active sellers already looking for tools. I pitched them: Give me 48 hours to build the solution. If you like it, we partner. No strings attached.

I built it, and they called the next morning and told me to quit my job.

Day 30, Launch Fast hit $10K MRR. Day 60, $17–18K. By Day 90, we were at $21.8K. All paying users, a Chrome extension with 330 active users, and a product built entirely in Cursor by a guy who still can't read a stack trace without AI help.

That was a few months ago. We're at $30K MRR now. We've been improving the core research engine, and we just shipped our MCP, one of the first all-in-one agentic tools for Amazon sellers. It lets them research products, manage their catalogs, and optimize ads through AI without bouncing between five different tools. That's landed well.

I have also started taking on contracts, helping other founders build their products A-Z, and running enterprise workshops on Claude Code and agentic workflows. Turns out "non-technical founder builds SaaS in 48 hours" opens a lot of doors.

Building under a self-imposed deadline

I built the initial product under a self-imposed deadline — the one that I told Legacy X. "Give me 48 hours."

I didn't say that because I was confident; I said it because I needed a forcing function, or I'd keep planning instead of shipping.

• The first four hours, I didn't touch Cursor at all. I mapped their existing workflows, SOPs, and the data they were already using with students. I needed to understand their processes before building something that fit. That became the foundation of the MVP.

• I spent hours five through twelve in Cursor building the core features. It didn't need to be perfect; it needed to be functional enough that someone could see what it would become.

• I spent hours 13 through 20 testing and iterating.

• I focused on the UI during hours 21 through 30, which a lot of first-time builders skip. I'd learned from running Amazon brands that branding is everything. If it looks unfinished, people assume it is.

• I spent hours 31 through 40 on edge case testing, ensuring nothing broke under real use.

• For the final stretch, I recorded the demo video and sent it over. That was the pitch.

Changing the stack

The stack was Cursor for the build, Vercel for hosting, Supabase for the database and auth, Resend for onboarding emails, Apify for data aggregation, TypeScript and Next.js throughout. The whole thing cost almost nothing to run until it started making money.

That's not the stack anymore.

We've moved a lot of infrastructure to Cloudflare and built our own crawlers and data layer engine from scratch. Apify did its job at the MVP stage; it let me move fast without building data infrastructure from scratch. Once we knew the product worked, we needed more control over how data flows through the system. Building our own was the right call.

For the build workflow, I've been running Superset as my terminal manager, using both Codex and Claude Code depending on the task. Having both in the same environment has changed how fast we ship.

Core frontend is still Next.js and React. That hasn't changed, and I don't see a reason to change it. The stack is well-documented, AI-friendly, and agents work well with it.

Partnerships, education, and ads

The business model is subscription-based. Legacy X coaching customers have their own plan. And we have public usage tiers at $49, $89, and $199 a month. All paying customers, no free tier.

The first users didn't come from marketing. They came from a partnership. Legacy X had thousands of active Amazon sellers already looking for better tools. I traded equity for access to that distribution. Day one, we had real customers. That's not a replicable hack for every founder, but it solved the hardest problem as quickly as possible.

From there, we had to build our own acquisition engine.

• Education became the core of it. We run weekly sessions teaching new sellers how to use the software, but also how to think about the Amazon business more broadly. People who feel you teach them tend to stay with you. That retention feeds the word of mouth, which feeds new signups inside seller communities where trust travels fast.

• On the content side, technical blog posts have built up Google traffic steadily over time. We also build free tools as a top-of-funnel play for sellers who are early in their journey and not ready to pay yet.

• Meta ads handle a portion of paid acquisition. We're also building out social this year, both through original content and agentic outbound on Reddit and LinkedIn.

The real growth driver, though, is product quality. Every time the core research engine improves meaningfully, we see it in referrals and retention. Sellers talk to each other constantly. If your tool does what it says it does, that community does a lot of the selling for you.

Start with systems, not tasks

The biggest challenge was something I didn't recognize as a problem until we were deep into it: I treated everything as a task instead of a system.

Early on, that's fine. You move fast, ship every day, and put out fires. Task mode works when the product is small, and you can hold all of it in your head. But as a founder, you cover tech, marketing, support, everything simultaneously. At some point, what got you to $10K MRR becomes the ceiling that keeps you from getting to $50K.

Moving toward agents forced me to see it. When you start building agentic workflows, you realize the architecture you should have built from day one. Observability pipelines, automated bug detection, and systems that flag and address issues without needing you in the loop. I built all of that reactively, as problems surfaced, instead of building it into the foundation.

If I started over, I would set up those pipelines before they were urgent. As a founder, the goal isn't to be the best at solving problems. It's to build systems that solve problems without your constant input. The product should get better even when you're not working on it.

That's what I'd do differently. I'm focused on building that now.

Focus on the domain and distribution

Three pieces of advice I'd tell anyone starting out:

• First, build where you already have domain knowledge. Not where you think there's an opportunity, not where you read that the market is big. Build for a problem you have personally experienced. I failed ten or twelve times before Launch Fast because I was building in spaces I didn't belong to. The one that worked was where I was the user. You probably have years of experience in something most developers don't. That knowledge is the product.

• Second, solve the distribution problem before worrying about everything else. The biggest mistake I see first-time builders make is assuming a great product will find its audience. It won't. Find someone who already has your target customers and figure out how to get in front of them. It might cost you equity. It might cost you a revenue share. Pay it. Fifty percent of something real is better than one hundred percent of nothing shipped.

• Third, use a forcing function to ship. When I built the Launch Fast MVP, I gave myself 48 hours and sent a demo to a potential partner. I didn't have a perfect product. I had something functional enough to show what it could become. The discipline of a real deadline with a real recipient changed everything. Set yours before you feel ready.

The tools exist to build without a CS degree. The knowledge to build something people need already exists in your head. The only thing standing between you and shipping is the decision to stop planning and start.

What's next?

Launch Fast is the core, and growing it is the priority. We're nowhere near the ceiling on what the product can do for Amazon sellers, and the agentic layer we are building is still early. I want to keep inventing new primitives specifically for this space — tools and workflows that sellers can't get anywhere else. The research engine is one piece. There's a lot more to build.

Outside of that, I'm genuinely interested in becoming a better agentic developer for its own sake. I constantly build random ideas. Most won't turn into businesses, and I'm fine with that. The reps matter. Every project teaches me how agents behave in production versus how they behave in demos, and that gap is where the interesting problems live.

The workshops and consulting work are becoming their own thing worth developing. Enterprise teams currently have a real demand to understand how to build with Claude Code and agentic workflows, but lack internal product-level expertise. By embedding myself as a forward-deployed AI consultant with those teams, I stay close to the hard problems, not just the theoretical ones.

Overall, I just want to keep building. Launch Fast, side projects, client work. My goal is to improve all of it simultaneously.

Follow along

You can find the product at launchfastlegacyx.com. If you sell on Amazon or are thinking about it, that's the place to start.

For development insights, I share what I'm working on and learning in real time on X. There, you'll find most of the agentic development content and random project updates.

To get in touch about consulting or enterprise workshops, visit my personal site: hasaamb.com.